How we test:
It's easy to take your display for granted. But when you consider that you probably spend more time using your display than any other component of your PC, you'll come to realize that the screen you stare at every day is a vital part of your computer system. An LCD's specs can only tell you so much; ultimately, what differentiates one LCD from another is image quality. To evaluate a display's performance, we use a variety of DisplayMate test screens in conjunction with diagnostic equipment and the most important tool, the human eye. Each test screen we use is specifically designed to emphasize a particular area of performance, such as text readability, color accuracy, or image uniformity.
CNET Labs' LCD test bed consists of a 3.4GHz Pentium 4 550 processor, 1GB of DDR SDRAM running at 533MHz, an Nvidia GeForce 8600 GTS
-based graphics card, a Sony BWU-100A
Blu-ray drive, and a 74GB Western Digital WD740 Raptor hard drive, running Windows XP Professional SP2. We test all LCD monitors with Windows' Display Properties set to 32-bit color and use the flat-panel display's native pixel resolution
as specified by the manufacturer. When doing side-by-side comparative testing, we use an Extron Electronics D2 DA4 DVI Direct Digital Distribution Amplifier to simultaneously send the video signal from our test bed out to as many as four LCDs.
Variables such as the light source, the viewing angle, and the system's graphics card can have a dramatic effect on an LCD's performance. To maintain consistent and ideal viewing conditions, all testing is performed in a controlled lighting environment using CinemaQuest Ideal-Lume
lights, which help to preserve accurate color perception. Additionally, each display is adjusted to perform optimally, based on the manufacturer-recommended or default settings when possible.
CNET Labs tests luminance levels (and whether these levels are in line with manufacturers' claims) using the Minolta CA-210 LCD color analyzer
. Because LCDs typically vary in brightness across the entire screen, we take brightness readings from different sections of the display as laid out in the VESA Brightness Uniformity Test screen found in the DisplayMate test suite. The brightness number that CNET publishes represents an average of the luminance readings taken at nine specific points on the screen.
To test an LCD's contrast ratio, we use the display's default contrast and brightness settings. Using the Minolta CA-210 LCD color analyzer
in a completely darkened room, we measure the brightness of both the light and dark squares found on the DisplayMate ANSI Checkerboard Contrast screen. We divide the average luminance of the white squares by the average luminance of the dark squares to yield the LCD's contrast ratio.
With guidance from DisplayMate Technologies
, we created our own scripted selection of test screens in DisplayMate Multimedia Edition
. These screens are designed to isolate common phenomena such as digital noise, streaking and ghosting, ringing and overshoot, and color-tracking errors.
We evaluate each display in four categories: sharpness, grayscale range, color quality, and image uniformity. We compile the product's performance scores from each category and distill them into a single performance rating: the CNET Labs DisplayMate test score.
The test screens described below represent our core LCD testing. Most of the screens in this test suite can be configured in a number of different ways, such as altering the background and foreground colors. Depending on the characteristics of an individual display, we might use several variations of these screens as well as additional DisplayMate screens not found in our custom script to conduct further testing.
Stuck-pixel test (DisplayMate screen 10:04)
LCDs sometimes have pixels that are either permanently shut off (dead) or permanently turned on (stuck in one color). This test consists of a series of colored screens that make bad pixels easy to spot. Most LCDs we test have a few bad pixels, but we check to see that each LCD has no more than the number deemed acceptable by its manufacturer.
Pixel tracking and phase-lock test (DisplayMate screen 10:16)
This screen helps reduce noise fluctuations and sharpens image quality by optimizing an LCD's pixel tracking and timing lock, which is sometimes referred to as phase, timing,
or horizontal time
Page of text (DisplayMate screens 7:1 and 7:2)
These screens illustrate a display's ability to render text under a variety of conditions. We cycle through various text and background colors, view split screens with inverse text and background colors, and adjust the type and the size of a font.
Video bandwidth index (DisplayMate screen 7:18)
This adjustable checkerboard pattern allows us to measure an LCD's video bandwidth. Perfect bandwidth indicates that broad and fine details are produced identically. A bandwidth that is too low or too high will lead to a loss of fine image detail. This screen also indicates whether the separate red, green, and blue color channels remain consistent in hue or the color shifts when moving from broad to fine image detail.
Intensity and grayscale (DisplayMate screen 9:30)
LCDs often have trouble reproducing all of the levels of the grayscale
(the range of grays between true black and true white). This screen helps to identify a display's ability to deliver seamless gradation across the full spectrum of grays, both horizontally and vertically across the screen.
Extreme grayscale bars (DisplayMate screen 9:25)
As the title suggests, this screen has the dual function of evaluating the darkest and brightest areas of the grayscale. These outermost edges are the most difficult part of the scale for LCDs to produce. We use this screen primarily for the dark end of the scale to check an LCD's ability to deliver a true black and still produce the darkest grays of the grayscale.
Low saturation colors (DisplayMate screen 9:23)
When producing a bright white image, many LCDs oversaturate the grayscale: the lightest grays of the scale are lost in the white background. Oversaturation can also lead to loss of color range; this screen is used to evaluate color reproduction at the brightest end of the scale, closest to white.
Color tracking (DisplayMate screen 9:20)
A color-tracking error occurs when the intensity of red, green, and blue (RGB) do not adjust identically with signal-level changes. This lack of balance between the RGB channels affects color as well as grayscale, but it is most easily identified as a shift in color within shades of gray. We use this screen to look for grays that appear to be tinted with color.
64-256 intensity color ramp (DisplayMate screen 9:32)
Similar to the grayscale tests, the color ramp illustrates an LCD's capacity to render gradations of primary colors smoothly, uniformly, and consistently. This screen is also used to check that the colors don't shift hue as the color levels increase or decrease.
256 intensity color ramp (DisplayMate screen 9:33)
Similar to the grayscale tests, the color ramp illustrates an LCD's capacity to render gradations of primary colors smoothly, uniformly, and consistently. This screen is also used to check that the colors don't shift hue as the color levels increase or decrease. This is a good screen to test for evidence of color banding.
Color scales (DisplayMate screen 9:41)
Similar to the intensity color ramp, the color-scales screen helps us evaluate the smooth gradation of colors, expanding the palette to 10 principal colors.
Screen uniformity (DisplayMate screen 10:01)
Perfectly uniform backlighting across an LCD's entire display surface is difficult to achieve. It is not uncommon for an LCD to have bright or dim patches or subtler variations in color intensity, which give the appearance of shading across the screen, or variable color intensity on the display. We use this screen to check for irregularities caused by backlighting issues or other screen-uniformity factors, such as variations or reflections inside the glass panel.
Dark screen (DisplayMate screen 10:19)
Because this test screen is designed to appear uniformly black, it is useful for evaluating an LCD's black-level capabilities. Additionally, a dark screen is the easiest way to spot glare and reflection problems, both of which can have distracting effects when you're viewing an LCD.
Streaking and ghosting (DisplayMate screen 10:42)
This screen helps us detect streaking
--light or dark shadows that trail an image in areas where large changes in contrast are present. This should not be confused with the streaking that is often found in moving images. This test deals only with problems that arise when a display renders large, chunky graphic elements, such as bar graphs or tiled arrangements of open windows.
Viewing angle (DisplayMate screen 9:30)
Display manufacturers use a variety of techniques to measure an LCD's viewing angle--making comparisons via the manufacturer's stated specs practically impossible. We've standardized on a methodology that permits us to make direct comparisons between different LCDs. With DisplayMate's Intensity and Grayscale test screen showing on the display, we tilt the display right, left, and back, measuring the angle at which the white portion of the grayscale begins to degrade.
In addition to our suite of DisplayMate test screens, we also use a number of multimedia tests designed to mirror real-world use, such as Blu-ray movie playback, DVD movie playback, and gameplay. For our Blu-ray and DVD playback tests, we use CyberLink's PowerDVD Ultra
in full-screen mode with a Swordfish
Blu-ray disc, Chapters 4, 6, and 10, and a Kill Bill Volume 1
DVD, Chapters 8, 15,
and 16, respectively. We use these movies to evaluate an LCD monitor's detail and sharpness, color quality, and streaking or ghosting problems. To evaluate detail and sharpness, we look at the amount of digital noise in the pictures, the display's ability to reproduce fine details in brighter and darker areas of the images, and the overall sharpness of the images. For color quality we evaluate the display's ability to reproduce bright, vivid, accurate colors, including bright whites and solid blacks. Streaking and ghosting are issues with most LCD monitors, so for our tests, we assess the severity of the streaking and ghosting and score accordingly. For games testing, we use World of Warcraft
and utilize many of the same evaluation methods that we use with movie testing.
CNET Labs amended its LCD monitor testing methodology in July 2007. As a result, the performance scores for LCD monitors reviewed before July 2007 cannot be directly compared with those tested after.